 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Java Forum / General / January 2006Catch outofmemory exception
Is it possible to catch java.lang.OutofMemoryException ? >Is it possible to catch java.lang.OutofMemoryException ? Yes. > >Is it possible to catch java.lang.OutofMemoryException ? > > Yes. go on.. >>>Is it possible to catch java.lang.OutofMemoryException ? >>Yes. >go on.. The correct spelling is "java.lang.OutOfMemoryError". So one would write try { /* ... */ } catch( final java.lang.OutOfMemoryError e ){ /* ... */ } > > >Is it possible to catch java.lang.OutofMemoryException ? > > > > Yes. > > go on.. No, because there's no such exception. There's OutOfMemoryError. You can attempt to catch it, though, since it's Throwable just as Exception is. Reading up on Java memory management issues, you'll find that it's almost never feasible to really catch it. By the time it's thrown, it's likely that the situation is beyond repair and your JVM is going to terminate.  SignatureSteve W. Jackson Montgomery, Alabama If there's a reasonable chance an OutOfMemoryError will occur, it may make sense indeed to catch it. Steve's point is true -- if this gets thrown, the situation is likely beyond repair. So the only practical purpose of catching this error is to die gracefully, i.e. with a friendly error message instead of an exception trace. >So the only practical purpose of catching this error is to die >gracefully, i.e. with a friendly error message instead of an >exception trace. Here is a block trying to allocate components into a pool as long as possible (as much as memory is available). Eventually about 30 % of the pool then will be released again to have some memory for other purposes. { char[] buffer = new char[ 65535 ]; try { while( true )pool.add( new Component() ); } catch( final java.lang.OutOfMemoryError error ) { buffer = null; java.lang.System.gc(); for( int i = 0; i <( int )( pool.size() * .3 ); ++i ) pool.removeLastComponent(); java.lang.System.gc(); }; } >> So the only practical purpose of catching this error is to die >> gracefully, i.e. with a friendly error message instead of an [quoted text clipped - 11 lines] > for( int i = 0; i <( int )( pool.size() * .3 ); ++i ) > pool.removeLastComponent(); java.lang.System.gc(); }; } Some other thread could allocate just a few bytes at the wrong time causing an OutOfMemoryError that you don't catch. You would probably have to provide a default uncaught exception handler (http://java.sun.com/j2se/1.5.0/docs/api/java/lang/Thread.html#setDefaultUncaught ExceptionHandler(java.lang.Thread.UncaughtExceptionHandler)) to deal with this possibility. Dan. SignatureDaniel Dyer http://www.dandyer.co.uk > Here is a block trying to allocate components into a pool as > long as possible (as much as memory is available). Eventually > about 30 % of the pool then will be released again to have > some memory for other purposes. This does not look like a good idea. For starters, it looks like you are using a Throwable to control the normal flow of your program. In other words, you are depending on the Throwable (in this case, OutOfMemoryError) to be thrown in order for your program to continue normally. This is a Bad Thing. > Is it possible to catch java.lang.OutofMemoryException ? <facetious> There ain't no such creature. </facetious> It is possible to catch java.lang.OutOfMemoryError (note: Error, not Exception), but it is a profoundly bad idea to try it. -- Adam Maass > > Is it possible to catch java.lang.OutofMemoryException ? > [quoted text clipped - 4 lines] > It is possible to catch java.lang.OutOfMemoryError (note: Error, not > Exception), but it is a profoundly bad idea to try it. If you're running a pretty GUI application then I'd agree - the OutOfMemoryError can occur in your GUI thread and then you're stuck. The only reason to catch it is to die gracefully. However, in a server the #1 rule is "Don't die", so it's far from a bad idea to try it. Alun Harford > "Adam Maass" <adam.nospam.maass@comcast.net> wrote in message >> > Is it possible to catch java.lang.OutofMemoryException ? [quoted text clipped - 10 lines] > only reason to catch it is to die gracefully. However, in a server the #1 > rule is "Don't die", so it's far from a bad idea to try it. The problem with trying to catch OutOfMemoryError is that it could occur *anywhere* there's an allocation (ignorning for the time being that the class also represents certain other hard limits imposed by the OS that have been exhausted -- threads and graphics handles comes to mind). In particular, in a multi-threaded application, we don't know that the thread that gets an OutOfMemoryError will have the knowledge necessary to release enough references to make things OK. My first instinct is that I'd rather have a server application die with a nice stack trace than limp along producing spurious results. I might be convinced otherwise for certain situations... -- Adam Maass "I'd rather have a server application die with a nice stack trace than limp along producing spurious results" Yes, and that's the point: how do you get this nice stack trace? OutOfMemoryError does not print a nice stack trace by defualt. If you want it, you're going to have to catch it, and manually dump it to a file (presuming you don't want to squirt it onto a screen and hope it's still there next day). Catching OutOfMemoryError's, as has been said on this thread, is almost invariably not an attempt to correct anything, just to perform precisely this kind of post-mortum data. Even for a GUI. Would you rather you application just die, or print, "Out of memory," and then die when you click, "OK?" At least with, "Out of memory," the user knows he should load fewer files/analysis-branches/whatever. .ed -- www.EdmundKirwan.com - Home of The Fractal Class Composition. Bruce Lee a écrit : > Is it possible to catch java.lang.OutofMemoryException ? I personnaly catch StackOverflowError without any problem. So catching Error is possible. But the OutOfMemoryError can be thrown by almost anything, and thinking that it could be throwned by an only method is wrong. Imagine that youre method consumes almost all memory, and then an other thread of the VM allocates a few bytes: the error will be thrown by the thread, not by you're method. So catching such errors may be very hasardous. SignatureAlexandre CARTAPANIS - Responsable Système et Réseau Email alexandre.cartapanis@macymed.fr Gsm. 06 72 07 51 55 Macymed SARL - 9 bvd Kraëmer 13014 Marseille France Tél. 04 91 48 31 58 - Fax. 04 91 02 36 47 Web http://www.macymed.fr - Email info@macymed.fr Bruce Lee a écrit : > Is it possible to catch java.lang.OutofMemoryException ? > >>I personnaly catch StackOverflowError without any problem. So catching >>Error is possible. >>But the OutOfMemoryError can be thrown by almost anything, and thinking >>that it could be throwned by an only method is wrong. Imagine that youre >>method consumes almost all memory, and then an other thread of the VM >>allocates a few bytes: the error will be thrown by the thread, not by >>you're method. So catching such errors may be very hasardous. -- That's the problem. It'd be difficult for me to pinpoint where the error might occur so try catching everything would be a pain. The workaround I'm using is to use 2 JVMs so in the event the program throws a OutOfMemoryError then the other one will fire it back up again. Free MagazinesGet these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...
|
Reply to this Message
|
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.